You are here: Home » Opinion » Columns
Business Standard

Devangshu Datta: Man and machines

Devangshu Datta  |  New Delhi 

The most obvious example of translating into normal civilian use is the Internet, which was launched as a project to develop communications protocols to survive wartime network disruptions. The (GPS) is another defence project, which has now become ubiquitous. Since 9/11, the “war on terror” has driven advances in biometrics, machine translation, speech recognition and so on.

The next big set of commercial adoptions could be driverless vehicles and robot “coolies”. The uses “packbots”, multi-legged creatures to carry large loads of munitions, food, medicine and so on in inhospitable terrain. These are autonomous and capable of navigating optimal routes.

Drones – pilot-less aircraft of various sizes – have become objects of hatred owing to their use in bombing. But they have many other potential utilities. For example, they can provide emergency medical services. The Israelis have experimental drone ambulances. These have faster reaction times and they can reach places inaccessible by road. Drones can administer emergency aid via remote-controlled supervision, and also evacuate patients.

The payoffs for autonomous driverless cars, or roadbots as they are sometimes known, would be large. Driverless parallel and reverse-parking is already a feature in many cars. So is cruise control.

Roadbots don’t have variable attention spans; they don’t suffer heart attacks, epileptic fits, or get drunk. In heavy traffic, they can maintain optimal braking distances and speeds, by using sensors. They can even brake and accelerate in unison like an army platoon, if they are in communication with other smart cars.

The early roadbot trials started with the Eureka project between 1987 and 1995. Munich University collaborated with to convert S Class Mercedes cars into roadbots. These could negotiate multi-lane highway traffic without much intervention. On average, the Eureka bots required human intervention once every 9 km. The record was 158 km driven at 175 kmph on an Autobahn without human intervention.

Admittedly, German drivers are known for near-robotic adherence to rules. The Eureka vehicles weren’t capable of handling the more chaotic traffic of normal urbania. Nor could they negotiate poorer roads, or off-road terrain.

Enormous advances have since been made. Driverless vehicles should soon be capable of handling both on-road and off-road traffic challenges with zero, or near-zero human intervention. Research is being driven by a series of civilian and military projects.

The hopes that around one-third of its ground supply vehicles would be autonomous by 2015. To that end, the (the US Defense Advanced Research Project Agency) Urban Challenge offers large prizes for driverless cars that can navigate long distances, on and off-road, across difficult environments within a given time, while adhering to traffic safety rules.

Various editions have held in various ghost towns and on military bases, with routes littered with obstacles. The test vehicles are accompanied by normal vehicles driven by stunt drivers to simulate traffic, and the routes include narrow tunnels and sharp turns. The last Challenge winner, a Carnegie Mellon adaptation of a called “Boss”, was a converted SUV with the driving style of an “aggressive soccer mom”.

While European military research doesn’t hold high-profile races, there is the Elrob (Euro land robot) expo where robotic vehicular technologies are demonstrated. Designs to carry out preset tasks such as recce-surveillance, perimeter guarding, transport, autonomous navigation and so on, have been developed.

In 2010, the Artificial Vision and Intelligent Systems Laboratory of the Unbiversity of Parma drove four electric vehicles across 13,000 km from Parma to Shanghai. The journey took three months across several countries. The trip involved almost no human intervention, except mainly route-mapping by the lead car, where GPS wasn’t available.

At some stage, vehicle licensing laws will have to catch up to allow more widespread use of autonomous vehicles. A near-autonomous or fully-autonomous vehicle is almost certainly safer than the average human driver, as well as being more fuel-efficient. Most accidents are caused by easily avoidable human errors and some by absurd risk-taking, which can be eliminated by roadbots.

Many robotic safety features have already crept into modern auto-design. Common examples are speed governors, electronic stability control (for anti-skid), pre-collision warning systems (sensors hit the brakes and inflate air bags when they detect an imminent crash), active cruise control (using radar to detect cars ahead), pedestrian detection (scanning for the unwary jaywalker) and autonomous steering (correcting lane drift). Many auto-majors have ongoing research and have already incorporated such features. The era of the road-bot may soon be at hand.